Ischemic stroke is a worldwide public health problem and one of the leading causes of death in humans. A role for excitotoxicity-mediated by glutamate receptors has stimulated intensive research for decades. This has led to the hope that compounds antagonizing the glutamate receptor function may be of clinical benefit in treating stroke. However, the antagonist therapy failed in stroke trials, in most cases because of a limited therapeutic window and severe side effects, caused by the essential requirement of glutamate receptor-mediated excitatory neurotransmission in maintaining normal brain function.
Glutamate is the principal excitatory neurotransmitter in the brain and is involved in numerous physiological functions and processes including neuronal circuit development, learning and memory, as well as with many neuropathological disorders, such as the neurotoxicity associated with stroke. Glutamate activates two major subfamilies of ligand-gated postsynaptic receptors: AMPA (α-amino-3-hydroxyl-5-methyl-4-isoxazolepropionic acid) receptor and NMDA (N-methyl-D-aspartate) receptor (1). AMPA receptors mediate most of the excitatory postsynaptic current at resting membrane potentials while NMDA receptors are critically important in producing a number of different forms of synaptic plasticity in AMPA receptor-mediated synaptic transmission (2). Glutamate accumulation, in pathological condition such as immediately after ischemia, results in extensive stimulation of its receptors which can be highly neurotoxic (3,4). NMDA receptor-mediated neurotoxicity is dependent on extracellular Ca2+ and thus may reflect a large amount of Ca2+ influx directly through the receptor-gated ion channels (3,4). Most models of ischemic neurodegeneration have focused on the putative role of NMDA receptor activation. However, use of NMDA antagonists in animal models of ischemia as well as in human clinical trials has not generally shown the anticipated robust efficacy (5), suggesting NMDA receptor over activation may not be the sole player in the glutamate receptor-mediated neurotoxicity. AMPA receptors has been tightly associated with the selective pattern of neuronal loss in certain identifiable subsets of neurons observed in transient forebrain ischemia (6-13). However, as most AMPA receptor channels are much less Ca2+ permeable, the mechanism linking AMPA receptor activation to neuronal cell death remains largely unknown.
Functional changes in AMPA receptors are most often attributed to phosphorylation and de-phosphorylation by PKA (cyclic AMP-dependent protein kinase), protein kinase C(PKC) and CaM kinase II (calcium-calmodulin kinase II) (14-18). Recently, a variety of intracellular proteins have been reported to bind directly to AMPA receptors (19-23). These proteins play important roles not only in receptor targeting or clustering, but also in the modulation of receptor activity and activation of signaling pathways. One recent study reports that an extracellular secreted protein NARP binds to the extracellular N-terminus (NT) of AMPA receptors and plays a role in the induction of AMPA receptor clustering (24). This contrasts with all other identified AMPA interacting proteins that bind to the intracellular carboxyl tail (CT) of the AMPA receptor subunits.
Molecular Biology and Functions of GAPDH:
GAPDH is a tetrameric protein (144 kDa) composed of four identical subunits (37 kDa). The monomer is about 333-335 amino acids long, and each monomer has binding sites for the substrate (glyceraldehyde-3-phosphate, G-3-P) and co-factor nicotinamide adenine dinucleotide (NAD+) (25-26). Residues 0-149 from N-termini comprise the NAD+ binding domain; and, side chains involved in catalysis are contained in residues from 149-333 or 149-335. The co-factor binds reversibly to the enzyme prior to the substrate binding.
Traditionally, GAPDH has been considered the key enzyme in glycolysis, with a critical role in energy production. It is considered to be the product of a housekeeping gene whose transcript level remains constant under most of experimental conditions. However, recent evidence supports the notion that GAPDH plays a critical role in apoptosis during which its expression and subcellular localization is altered (27-30). The cellular localization of GAPDH is not only restricted to the cytosol but it is also found in the nucleus and plasma membrane.
In the nucleus, GAPDH has been shown to act as a DNA binding protein and t-RNA transport protein which plays a specific role in the transportation and maintenance of nucleic acid. GAPDH binds to and transports t-RNA from the nucleus to the cytosol, and the interaction of GAPDH with t-RNA is displaced by the co-factor, NAD+(31-32). In addition, the uracil DNA glycosylase activity of GAPDH, together with its binding to diadenosine tetraphosphate (Ap4A), imply that nuclear GAPDH is involved in DNA replication and repair (33).
In the cytosol, RNA/GAPDH interactions enable GAPDH to play an important role in translational regulation of gene expression by controlling rate of protein synthesis and/or by altering the stability of mRNA (34-35). Furthermore, GAPDH is essential for ER to Golgi transport through its interaction with Rab2 GTPase and atypical protein kinase C/(aPKC/), two important proteins involved in the early secretory pathway and vesicle formation (36-38).
The function of membrane-associated GAPDH is to bind to tubulin thereby regulating polymerization and bundling of microtubules near the cell membrane, suggesting that GAPDH is involved in the re-organization of sub-cellular organelles (39). Furthermore, release of tubulin from membrane-associated GAPDH facilitates the fusion of vesicles to the plasma membrane (40). Thus, GAPDH is involved in both maintenance of membrane trafficking and the promotion of vesicle fusion through modulation of cytoskeleton functions.
GAPDH and Apoptosis:
GAPDH is overexpressed and accumulated in the nucleus during apoptosis induced by a variety of insults. Evidence shows that the GAPDH nuclear translocation is essential for the apoptotic cascade (41-42). Western blot analysis and confocal immunocytochemistry results indicate a significant increase of GAPDH expression in the nuclear fraction subjected to various stresses. Antisense oligonucleotides that deplete GAPDH prevent this nuclear translocation and reduce apoptosis (41, 43-44).
The mechanism underlying GAPDH nuclear translocation and subsequent cell death remains largely unknown, however, recent studies have suggested several potential factors/pathways that maybe involved in the process: the expression of GAPDH is regulated by p53, the tumor suppressor protein and by proapoptotic transcription factor. Thus, GAPDH could be one of the downstream apoptotic mediators (45); over expression of bcl-2 blocks the apoptotic insults triggered by GAPDH over expression, nuclear translocation and subsequent apoptosis, suggesting that Bcl-2 may participate in the regulation of GAPDH nuclear translocation. This effect may be part of the mechanism of Bcl-2-induced protection against apoptosis (46) and GAPDH binds to a nuclear localization signal containing protein, Siah which initiates its translocation to the nucleus. The association with GAPDH stabilizes Siah and thereby enhances Siah-mediated proteolytic cleavage of its nuclear substrates, such as N-CoR and triggers apoptosis (44, 47-49).
Molecular Biology of AMPA Receptors:
AMPA receptors are intrinsic ion channels comprised of different subunits, which are encoded by four gene products, termed GluR1, 2, 3 and 4 (50-54). AMPA receptors are believed to exist as heteromeric assemblies of these subunits. Each subunit posses an extracellular NT domain, four putative transmembrane (TM) domains of which the second is believed to be a reentrant loop, as well as an intracellular CT domain (55-56). It is thought that the M2 loop participates in the formation of the ion channel pore. Two 150 amino-acid sequences (termed as S1 and S2) which are separated by the M1-M3 membrane domains appear to represent the agonist recognition sites (57). The molecular determinant of the calcium permeability is localized to the single amino acid in TM 2 region. A positively charged arginine (R) residue is found in position 586 for GluR2 whereas a neutral glutamine (Q) is found in the same position of GluR1, GluR3 and GluR4 subunits. Recombinant AMPA receptors lacking GluR2 show high calcium permeability and current-voltage relationships that doubly rectify (58). All four AMPA receptor subunits occur in two alternatively spliced versions, flip and flop. Flip differs flop version in the profile of desensitization and these variants show differing regional distributions which vary during development (59-60). The exact subunit composition of native AMPA receptors is not clear, but immunoprecipitation strategies have shown two major complexes composed of GluR2 together with either GluR1 or GluR3 in rat hippocampus (61). The presence of GluR2 subunit greatly reduces Ca++ and Zn++ permeability (58, 62-65) as well as the single channel conductance (66) of these receptors. Hence, most of AMPA receptors at the hippocampal synapses are Ca++ and Zn++-impermeable (62, 67-68).
AMPA receptor interacting proteins and their function: Using yeast a two-hybrid system with the CT domain of GluR2 subunit as bait, GRIP (Glutamate Receptor Interacting Protein, also known as AMPA receptor-binding protein, ABP) was the first protein identified as an AMPA receptor interacting protein (20). This finding was followed by extensive efforts to identify other AMPA receptor interacting proteins. Ban 4.1 and PKCy interact with both GluR1 and GluR4 subunits (69-70); SAP97 (synapse-associated protein-97) couples only with GluR171; GRIP1, 2 and PICK1 (protein interacting with C kinase) bind to GluR3 and GluR4c (19,77). Also, three additional proteins, Stargazin, NARP (neuronal activity-regulated pentraxin), and AP2 (adaptor protein-2) bind to all of the AMPA receptor subunits (24, 72-73).
Interactions with the GluR2 subunit of AMPA receptors are of considerable interest due to the key biophysical properties conferred by the presence of this subunit. GRIP1, 2, PICK1, and NSF (N-ethylmaleimide-sensitive factor) are identified as GluR2 interacting proteins (20-21, 74-77). Two distinct interaction domains have been identified for the GluR2 C-terminus. NSF protein binds to a more proximal site (74,76), while the proteins GRIP1, ABP, and PICK1 associate with the PDZ-binding motif at the very distal end of the C-terminus (19-20, 76).
AMPA receptor interacting proteins may regulate these receptors in a variety of ways, such as altering AMPA receptor localization, clustering and/or trafficking. The binding of GluR2/3/4 to PICK1 is involved in the clustering of AMPA receptors (19,77), while the binding of GluR2/4 with NSF likely regulates rapid turnover of synaptic receptors (21, 74-75). Disruption of GluR2/3-GRIP interactions causes an increase in synaptic currents and prevents the generation of LTD22 and interaction with F— actin also plays a role in location of AMPA receptor clusters (78).
GluR2 subunit trafficking: Understanding the mechanism controlling surface expression of AMPA receptors in insult-vulnerable neurons is important because 98% of these receptors are localized at the synapse (hippocampus) (79-80) and the modulation of membrane receptor expression is an efficient mechanism for regulating the efficacy of synaptic transmission (80-98). AMPA receptors are trafficked between the plasma membrane and the intracellular compartments via delivery (insertion) and internalization (endocytosis) pathways. Native AMPA receptors undergo clathrin-dependent constitutive and regulated internalization involving adaptor protein-2 (AP2) and dynamin (99-100). Constitutive internalization counteracts constitutive receptor insertion, ensuring a constant number of cell surface AMPA receptors. Both receptor phosphorylation and GluR2 interacting proteins play an important role in trafficking of these receptors. Furthermore, NMDA receptor activity can regulated both AMPA receptor membrane insertion and internalization and this is important in certain forms of synaptic plasticity (100) as well as in NMDA-mediated neuronal apoptosis (101).
Glutamate mediated neurotoxicity is thought to contribute to neurodegeneration following a wide range of neurological insults including ischemia, trauma, hypoglycemia and epileptic seizure (3,4). It is believed that elevation of the extracellular glutamate after cerebral ischemia plays a critical role in the patho-physiological processes leading to death of ischemic brain tissue (102-103). Excessive glutamate, through an action on mainly on NMDA and AMPA glutamate receptors, facilitates Ca2+ influx, which under pathological conditions can result in excitotoxicity. The “calcium overload” hypothesis is the prominent theory explaining excitotoxicity (4). The molecular mechanisms underlying NMDA-mediated excitotoxicity involve many Ca2+-regulated processes in the cell including activation of proteases (104), endonucleases (105), nitric oxide synthase (106), the production of free radicals (107) and mitochondrial membrane permeability (108). The “calcium theory” can also apply to the Ca2+ permeable AMPA receptor-induced toxicity, however, there must be another explanation for the Ca2+-inpermeable AMPA receptor induced toxicity. One possibility for Ca2+-impermeable AMPA receptor induced toxicity is to induce membrane depolarization via Na+ influx. The AMPA-mediated depolarization, in turn, opened both VSCCs and removed the Mg2+ block from NMDA receptors, thus allowing Ca2+ influx through these pathways (109-110). Another possibility is that AMPA receptor-mediated ion fluxes could be coupled to downstream neurotoxic second messengers via interactions with submembrane proteins. For example, the interaction of GRIP1 with GRASP-1 may couple AMPA receptors to Ras signaling (111) and GRASP-1 has been shown to be a neuronal substrate for caspase-3 (111) which is cleaved in apoptotic neurons in a time-dependent manner during development and ischemia (112). Furthermore, the potential role of GluR2-interacting proteins in excitotoxicity may be that the presence of GluR2 is required to maintain synaptic structure and organization. Accordingly, the toxicity observed in GluR2-deficient neurons may result from the effects on synaptic organization and function rather than due to AMPA receptor Ca2+ permeability. An interesting candidate protein is the NSF, as it has been shown both to interact with GluR2 and to mediate membrane-fusion events (113-115). Interestingly, NSF expression increases following an ischemic insult (116). It is not yet clear whether an increase in NSF leads to an increase of surface expression of existing GluR2-containing AMPA receptors following ischemia. If so, one may speculate that increased GluR2 surface expression may decrease Ca2+ permeability through AMPA receptors, and restore synaptic organization. Taken together, these activities indicate AMPA receptor interacting protein may play an important role in AMPA receptor-medited neurotoxicity.
The “GluR2 hypothesis” in AMPA receptor-mediated neurotoxicity (117-121) predicts that a relative reduction in the expression of GluR2 results in enhanced Ca2+-influx through newly synthesized AMPA receptors, thereby increasing neurotoxicity; and enhancing GluR2 membrane expression may provide protective effect based on the evidence showing that: (1) in ischemic CA1 neurons AMPA receptor-mediated EPSCs show an increased sensitivity to N-(4-hydroxyphenylpropanoyl)-spermine (NHPP-spermine) (122-123), a selective blocker for GluR2-lacking AMPA receptors (124-125). Indicative of a reduction in the number of GluR2 containing receptors; ischemic insults promote internalization of GluR2-containing AMPA receptors from synaptic sites and facilitate delivery of GluR2-lacking AMPA receptor (126); GluR2 expression is down regulated in vulnerable neurons in animal models of transient forebrain ischemia and epilepsy (127) and vulnerable CA1 pyramidal neurons can be rescued from forebrain ischemic injury by enhancing the expression of GluR2 containing receptors (127-128).
This evidence indicates the role of GluR2 membrane expression in the AMPA receptor-mediated neurotoxicity, which raise the possibility for proteins that regulate GluR2 subunit trafficking through protein-protein interaction with GluR2 to be involved in the AMPA receptor mediated apoptosis.
There is a need in the art for compositions and methods for modulating AMPA receptor-mediated excitotoxicity. There is also a need in the art for compositions and methods for modulating GAPDH association with GluR2 subunit or p53.